A Study of the Effect of Vaccinated Populations on the Prediction Accuracy of the Regression Models for Daily Covid-19 Cases Prediction

Covid-19 is a major health risk for the world. Almost all the countries got affected by this disease. It is an infectious disease and people can be affected easily. Serious patients require hospitalization and intensive care. Various vaccines have been developed to prevent the spread of Covid-19 Due to limited medical resources, it is important to know future daily new Covid-19 cases in advance. Regression methods have been used to predict future daily new Covid-19 cases. Almost all the studies in this area were carried out with data when the population was unvaccinated. It is expected that the Covid-19 spread will be different when significant population are vaccinated. In this paper, we study the effect of the vaccinated population on the accuracy of regression methods. We investigate an important question, should we use complete data or only the recent data (when the significant population are vaccinated) for the regression methods. We carried out studies with data from five countries UAE, UK, Singapore, Germany, and France. The results suggest that results are different with different data. For four countries, regression methods perform better with complete data. However, for UAE, regression methods perform better with the recent data. This suggests that it is important to do the experiments to select appropriate data for better prediction accuracy.